Mapping Scotland’s Drug Death Crisis: An Analysis of Trends, Disparities, and Policy Failures

Abstract

Executive Summary

This report examines drug misuse trends in Scotland from 1996 to 2023, highlighting key findings:

• Trends: Drug misuse deaths peaked in 2020 (1,339 deaths) but remained high at 1,276 in 2023.
  
• Causes: Accidental poisoning is now the leading cause of death, with poly-drug use contributing to over 70% of fatalities.
  
• Disparities: Individuals in the most deprived quintile experience drug death rates up to 8 times higher than those in the least deprived quintile.
  
• Policy Failures: Governance issues and inequitable funding exacerbate the crisis, despite targeted harm reduction measures like naloxone distribution.

---

Key Recommendations

1. Expand harm reduction services, including supervised injection sites and mobile units.
   
2. Scale trauma-informed, gender-sensitive services to address systemic inequities.
   
3. Integrate addiction and mental health care, focusing on deprived areas.

---

Introduction

The drug misuse crisis in Scotland remains one of the most severe public health challenges in Europe (McAuley, Matheson, and Robertson, 2022).

According to Scotland’s National Records, substance abuse caused 1,276 recorded deaths in 2023. At 22.4 fatalities per 100,000, Scotland’s drug death rate far exceeds the UK’s average of 7.6, reflecting ongoing systemic failures despite recent declines.

This paper examines drug misuse trends in Scotland from 1996 to 2023, focusing on:
- Geographic, socioeconomic, and gender disparities
- Policy interventions (e.g., the 2011 reform of reporting standards)
- The role of poly-drug use in fatalities

The analysis identifies temporal trends, disparities, and policy implications while offering actionable recommendations to reduce drug-related deaths in Scotland.

---

How to Use This Report

This report integrates interactive plots and maps to explore Scotland’s drug misuse crisis:

• Hover over data points for additional information.
  
• Use zoom and pan features in visualizations to examine specific time periods or categories.
  
• Click on lines in the legend to hide them from the graph, making it easier to focus on specific data trends.

---

Methods

This report utilizes data from Scotland’s National Records (1996–2023), integrating demographic, geographic, and substance-related factors.

Data cleaning, visualization, and statistical analysis were conducted using:
- ggplot2
- dplyr
- plotly

Interactive elements were added to enhance user engagement with the data.

---

Results and Analysis

This section explores Scotland’s drug misuse crisis through:
1. Statistical trends
2. Substance group analysis
3. Key disparities

Setting Up Environment

Load Libraries

# Load necessary libraries
# Each of these libraries is loaded to perform specific tasks in data manipulation, visualization, and handling various file formats.

library(readxl)   # Allows reading Excel files (.xls, .xlsx)
library(dplyr)    # Provides functions for data manipulation (filtering, summarizing, joining datasets)
library(ggplot2)  # Used for creating static visualizations (plots, bar charts, etc.)
library(janitor)  # Useful for cleaning column names and other basic data cleaning tasks
library(here)     # Helps manage file paths relative to the project's directory (avoids hard-coding paths)
library(tidyr)    # Provides functions to tidy the data (pivoting, reshaping, etc.)
library(openxlsx) # Used for reading and writing Excel files with more advanced formatting options
library(sf)       # Used for handling and analyzing spatial data (especially for working with shapefiles)
library(plotly)   # Used to create interactive plots (especially for web-based reports)
library(leaflet)  # Used for creating interactive maps (useful for geographic data visualization)
library(htmltools) # Facilitates working with HTML elements in R (for embedding HTML into R outputs)
library(RColorBrewer) # Provides color palettes for visualizations, helping make plots more visually appealing

Set Up File Path

# Set the file path to the Excel data file
file_path <- "C:/Users/alexa/OneDrive/Documents/year 4/forensics/essay/drug-related-deaths-scotland/drug-related-deaths-23-data.xlsx"

Trends in Drug Misuse Over Time

Drug-related deaths (DRDs) in Scotland rose from 244 in 1996 to a peak of 1,339 in 2020, before declining slightly to 1,276 in 2023 (Figure 1). These trends reflect the evolving challenges of tackling drug misuse, especially the impact of benzodiazepines and opioids.

The COVID-19 pandemic exacerbated vulnerabilities by disrupting support services, increasing social isolation, and tripling loneliness for those facing mental health challenges (O’Sullivan et al., 2021).

---

Key Takeaways:

• Benzodiazepines: Sharp rise post-2015, particularly street benzos.
  
• Opioids: Persistent contributors to drug misuse mortality.
  
• Policy Influence: The 2011 Naloxone Programme showed measurable impact.
  
• Pandemic Impact: Increased isolation amplified risks.

Figure 1: Trends in Drug Misuse Deaths in Scotland by Gender (1996–2023)

• The dashed navy line marks a notable policy change in 2011, which corresponds to the implementation of the National Naloxone Programme.
  
• Male deaths consistently exceeded female deaths, highlighting gender disparities in drug misuse trends.
  
• The peak in total deaths occurred in 2020, with both genders contributing to the surge.
  
• Post-2020, a slight decline in drug-related deaths has been observed, but the numbers remain alarmingly high.

Key takeaway: Gender-specific vulnerabilities must be addressed to tackle the crisis effectively. While men are disproportionately affected, the systemic barriers faced by women—such as stigma and caregiving stress—must also be addressed.

# Load and process data: Read the Excel file, clean column names, and select relevant columns
gender_all_data <- read.xlsx(file_path, sheet = "Table_1", startRow = 6) %>%
  clean_names() %>%
  select(year, sex, drug_misuse_deaths) %>%
  rename(`Total Deaths` = drug_misuse_deaths)

# Create policy line data: Set the year 2011 and generate y-values for the dashed line
policy_line <- data.frame(
  x = rep(2011, 100),  # Set x to 2011 for policy year, to highlight policy change
  y = seq(0, max(gender_all_data$`Total Deaths`), length.out = 100),  # Create y-values, creates a smooth line for the policy change
  text = "Policy Change: A notable policy change occurred in 2011."  # Add tooltip text, to provide information for the plot
)

# Create the plot: Use ggplot to visualize the total deaths by gender and add the policy line
gender_all_plot <- ggplot(gender_all_data) +
  geom_line(aes(x = year, y = `Total Deaths`, color = sex, group = sex), linewidth = 1.2) +  # Plot lines for gender, to visualize trends by gender
  geom_point(aes(x = year, y = `Total Deaths`, color = sex), size = 3) +  # Add points to the line, making trends more visible
  geom_line(data = policy_line, aes(x = x, y = y, text = text), color = "navy", linetype = "dashed", size = 1) +  # Add the policy line, to mark the policy change in 2011
  labs(
    title = "Drug Misuse Deaths in Scotland (1996–2023)",
    x = "Year",
    y = "Total Deaths",
    color = "Gender"
  ) +
  scale_color_manual(
    values = c("Females" = "#66c2a5", "Males" = "#8da0cb", "Persons" = "#fc8d62")  # Custom colors, to distinguish between genders
  ) +
  theme_minimal() +
  theme(
    legend.position = "bottom",  # Position the legend, to keep it out of the way
    legend.title = element_text(size = 12, face = "bold"),  # Style legend title, improving readability
    legend.text = element_text(size = 10),  # Style legend text, improving readability
    plot.title = element_text(size = 16, face = "bold"),  # Style plot title, to make it stand out
    axis.title = element_text(size = 12),
    axis.text = element_text(size = 10)
  )

# Convert ggplot to Plotly for interactivity and add the policy line as a separate trace
interactive_gender_all_plot <- ggplotly(gender_all_plot, tooltip = c("x", "y", "color")) %>%
  plotly::add_trace(
    x = policy_line$x,
    y = policy_line$y,
    mode = "lines",
    line = list(dash = "dash", color = "navy", width = 2),
    text = policy_line$text,
    hoverinfo = "text",  # Display tooltip text, to provide context on the policy change
    showlegend = FALSE  # Exclude the policy line from the legend. As it’s a static line, not part of the data
  )

# Display the interactive plot
interactive_gender_all_plot

Note: The policy change in 2011 aligns with the introduction of Scotland’s National Naloxone Programme, a critical intervention in reducing opioid-related fatalities.

Figure 2: Trends in Drug Misuse Deaths in Scotland by Substance Group (1996–2023)

• Benzodiazepines showed a dramatic rise post-2015, becoming the leading contributor to drug misuse deaths by 2020.
  
• Opioid-related deaths have shown persistent increases, further driving Scotland’s alarming drug death rates.
  
• Deaths involving stimulants and other substances remained relatively lower but have exhibited steady growth over the years.

Key takeaway: The emergence of high-potency illicit substances such as ‘street benzos’ highlights the importance of tailored interventions and enhanced harm reduction strategies to mitigate the crisis.

# Load and clean the data: Read the data from Excel and clean column names
table_3 <- read.xlsx(file_path, sheet = "Table_3", startRow = 6) %>%
  clean_names()

# Convert Year to numeric and clean substance columns: Remove non-numeric characters and convert columns to numeric
table_3_clean <- table_3 %>%
  mutate(across(
    c("any_opiate_or_opioid":"alcohol"), 
    ~ as.numeric(gsub("[^0-9.]", "", .))  # Remove non-numeric characters, to clean data for analysis
  )) %>%
  mutate(year = as.numeric(gsub("[^0-9]", "", year))) %>%
  filter(!is.na(year))  # Remove rows without valid year, to keep only valid data

# Reshape into long format and classify substances: Pivot data to long format and classify substance groups
table_3_long <- table_3_clean %>%
  pivot_longer(
    cols = c("any_opiate_or_opioid":"alcohol"),
    names_to = "substance",  # Pivot columns into key-value pairs, to facilitate easier grouping and analysis
    values_to = "deaths"
  ) %>%
  mutate(
    deaths = replace_na(as.numeric(deaths), 0),  # Replace missing deaths with 0, to clean data for analysis
    group = case_when(  # Classify substances into groups, to categorize data for visualization
      substance %in% c("heroin_morphine_note_6", "methadone", "buprenorphine", 
                       "codeine_or_a_codeine_containing_compound", 
                       "dihydrocodeine_or_a_d_h_c_containing_compound") ~ "Opioids",
      substance %in% c("any_benzodiazepine", "any_prescribable_benzodiazepine_note_7", 
                       "diazepam_note_8", "any_street_benzodiazepine_note_7",
                       "etizolam_note_8", "bromazolam_note_5_note_8") ~ "Benzodiazepines",
      substance %in% c("cocaine", "ecstasy_type", "amphetamines") ~ "Stimulants",
      substance %in% c("gabapentin_and_or_pregabalin", "alcohol", "nitazenes_note_5") ~ "Others",
      TRUE ~ "Unclassified"
    )
  ) %>%
  filter(!substance %in% c("all_drug_misuse_deaths", "single_drug_deaths", "poly_drug_deaths"),
         group != "Unclassified")  # Remove unwanted substances, to clean and focus the data

# Summarize deaths by year and group: Calculate total deaths for each group per year
filtered_data_clean <- table_3_long %>%
  group_by(year, group) %>%
  summarize(total_deaths = sum(deaths, na.rm = TRUE), .groups = "drop") %>%
  rename(
    Year = year,  # Rename year for consistency, makes it more readable for the plot
    `Total Deaths` = total_deaths  # Rename total_deaths, to improve clarity for plotting
  )

# Create a faceted plot: Plot total deaths by substance group across years
grouped_facet_plot <- ggplot(filtered_data_clean, aes(x = Year, y = `Total Deaths`)) +
  geom_line(aes(color = group, group = group), size = 1) +  # Add lines for each group, to show trends by group
  geom_point(aes(color = group), size = 2) +  # Add points, emphasizing data points for clarity
  labs(
    title = "Trends in Drug Misuse Deaths by Substance Group (1996–2023)",
    x = "Year",
    y = "Number of Deaths"
  ) +
  theme_minimal() +
  theme(
    legend.position = "none",  # Remove legend, to reduce clutter in the plot
    plot.title = element_text(size = 16, face = "bold"),
    axis.title = element_text(size = 12),
    axis.text = element_text(size = 10),
    strip.text = element_text(face = "bold", size = 12)  # Style facet titles, to make them stand out
  ) +
  scale_color_brewer(palette = "Set2") +  # Use color palette, to make the plot more visually appealing
  facet_wrap(~ group)  # Create facets by group, to compare trends across substance groups

# Convert to interactive Plotly plot
interactive_facet_plot <- ggplotly(grouped_facet_plot, tooltip = c("Year", "Total Deaths", "group")) %>%
  layout(
    title = list(text = "Trends in Drug Misuse Deaths by Substance Group (1996–2023)"),
    hoverlabel = list(font = list(size = 12))  # Customize hover label, to improve hover text readability
  )

# Display the interactive plot
interactive_facet_plot

In 2022, men accounted for 73% of drug-related deaths (DRDs), reflecting higher engagement in intravenous and poly-drug use (Gossop et al., 1994; Gjersing and Helle, 2021). The rise of ‘street benzos,’ characterised by high potency and unknown composition, has disproportionately impacted male users (McAuley et al., 2022).

Women, however, face distinct vulnerabilities, including:
- Caregiving stress
- Domestic abuse
- Stigma, compounded by welfare system changes (Tweed et al., 2022)

These disparities demand gender-sensitive harm reduction strategies that address structural inequities.

The 2011 policy change, which introduced new reporting procedures, marks a pivotal point in the timeline. Scotland implemented a National Naloxone Programme in 2011, which was associated with a 36% reduction in opioid-related deaths within four weeks of prison release.
- Specifically, the proportion of prison-release opioid-related deaths fell from 9.8% pre-2011 to 6.3% by 2013, demonstrating the program’s significant impact (Bird et al., 2016).

However, general practitioners showed limited awareness and involvement in naloxone distribution, citing:
1. A need for evidence of effectiveness
2. The necessity for appropriate training (Matheson et al., 2014)

These findings highlight the complex interplay of socioeconomic factors, healthcare services, and targeted interventions in addressing Scotland’s DRD crisis. They emphasise the importance of both structural changes and specific measures like naloxone distribution in comprehensive drug policy.

Critically, the persistent rise in deaths prior to 2011 underscores Scotland’s failure to address root causes such as:
- Socio-economic deprivation
- Inadequate mental health services

This trend highlights the need for proactive policies that address structural inequities alongside reactive measures like naloxone distribution.

---

Causes of Drug Misuse Deaths

Accidental poisoning emerged as the leading cause of death post-2011 (Figure 3), driven by the rising prevalence of synthetic opioids, which are 50–100 times more potent than morphine (Thompson et al., 2021).

These substances, often combined with benzodiazepines, are implicated in over 60% of drug-related deaths, highlighting the dangers of poly-drug use (McAuley et al., 2022).

Key takeaway: This increase reflects not only shifts in drug supply but also systemic shortcomings in prevention and intervention strategies.

---

Figure 3: Trends in Drug Misuse Deaths by Cause Over Time (1996–2023)

• The dashed line marks the division between pre- and post-2011 periods, following changes in reporting practices.
  
• Accidental poisoning has become the leading cause of death post-2011, while deaths attributed to mental disorders have decreased significantly, likely reflecting shifts in classification methods.

# Load and clean the data: Read and clean column names, rename for simplicity
table_2 <- read.xlsx(file_path, sheet = "Table_2", startRow = 6) %>%
  clean_names() %>%
  rename(  # Simplify column names for easier reference
    all_drug_misuse_deaths = "all_drug_misuse_deaths", 
    mental_disorders_due_to_drugs = "mental_and_behavioural_disorders_due_to_drug_use", 
    accidental_poisoning = "accidental_poisoning", 
    intentional_self_poisoning = "intentional_self_poisoning", 
    assault_by_drugs = "assault_by_drugs_etc", 
    undetermined_intent = "undetermined_intent"
  ) %>%
  mutate(year = as.numeric(year)) %>%
  filter(!is.na(year))  # Clean 'year' column, removing rows with invalid year data

# Reshape the data into long format and classify cause of death
table_2_long <- table_2 %>%
  pivot_longer(  # Reshape data, to make it easier to analyze across causes of death
    cols = c("mental_disorders_due_to_drugs", "accidental_poisoning", 
             "intentional_self_poisoning", "assault_by_drugs", 
             "undetermined_intent"), 
    names_to = "cause_of_death", 
    values_to = "deaths"
  ) %>%
  mutate(
    deaths = as.numeric(deaths),  # Convert deaths to numeric, to clean up the deaths data
    deaths = replace_na(deaths, 0),  # Replace missing deaths with 0, filling missing data
    period = ifelse(year < 2011, "Pre-2011", "Post-2011"),  # Categorize data by period, splitting data into pre and post-policy change
    cause_of_death = recode(  # Re-code causes of death, to make the categories easier to understand
      cause_of_death,
      "mental_disorders_due_to_drugs" = "Mental Disorders",
      "accidental_poisoning" = "Accidental Poisoning",
      "intentional_self_poisoning" = "Self-Poisoning",
      "assault_by_drugs" = "Assault",
      "undetermined_intent" = "Undetermined"
    ),
    hover_text = paste(  # Create custom hover text for interactivity, providing more detailed info on hover
      "Year:", year,
      "<br>Deaths:", deaths,
      "<br>Cause of Death:", cause_of_death
    )
  )

# Create a high-resolution data frame for the policy line (2011 policy change)
policy_line <- data.frame(
  x = rep(2011, 100),  # Fix x to 2011 for the policy line, marking the policy change year
  y = seq(0, max(table_2_long$deaths, na.rm = TRUE), length.out = 100),  # Create y-values, making the line smooth
  text = "Policy Change: A notable policy change occurred in 2011."  # Add tooltip text, to provide context for the policy line
)

# Create the base plot with ggplot, including the policy line
combined_plot <- ggplot() +
  geom_line(data = table_2_long, aes(x = year, y = deaths, color = cause_of_death, group = cause_of_death, text = hover_text), size = 1) +  # Add lines for causes of death to visualize trends
  geom_point(data = table_2_long, aes(x = year, y = deaths, color = cause_of_death, text = hover_text), size = 2) +  # Add points for data points, making data more visible
  geom_line(data = policy_line, aes(x = x, y = y), linetype = "dashed", color = "navy", size = 0.8) +  # Add dashed policy line,to indicate policy change
  labs(
    title = "Drug Misuse Deaths by Cause Over Time (1996–2023)",
    subtitle = "Dashed line marks the division between Pre- and Post-2011 periods.",
    x = "Year",
    y = "Number of Deaths",
    color = "Cause of Death"
  ) +
  theme_minimal() +
  theme(
    legend.position = "bottom",  # Place the legend at the bottom, to keep it out of the way
    plot.title = element_text(size = 16, face = "bold"),  # Style title, to make it stand out  
    axis.title = element_text(size = 12),
    axis.text = element_text(size = 10)
  ) +
  scale_color_brewer(palette = "Set2")  # Use color palette, to improve plot readability

# Convert to interactive Plotly plot
interactive_combined_plot <- ggplotly(combined_plot, tooltip = c("text"))

# Add the red dashed line trace for the policy change to the interactive plot
red_line_trace <- list(
  x = policy_line$x,
  y = policy_line$y,
  mode = "lines",
  line = list(dash = "dash", color = "navy", width = 2),
  text = policy_line$text,
  hoverinfo = "text",  # Show tooltip text, to display details on hover
  showlegend = FALSE  # Exclude the policy line from the legend, as not part of the data itself
)

# Add the red dashed line trace to the plot
interactive_combined_plot <- interactive_combined_plot %>%
  plotly::add_trace(
    x = red_line_trace$x,
    y = red_line_trace$y,
    mode = red_line_trace$mode,
    line = red_line_trace$line,
    text = red_line_trace$text,
    hoverinfo = red_line_trace$hoverinfo,
    showlegend = red_line_trace$showlegend
  )

# View the interactive plot
interactive_combined_plot

Note: The significant rise in accidental poisoning deaths reflects the growing impact of synthetic opioids and poly-drug use in Scotland’s overdose crisis.

---

Figure 4: Drug Misuse Deaths Over Time by Cause (Pre- and Post-2011)

The navy dashed line indicates the 2011 policy change impacting reporting practices.
- Accidental poisoning became the leading cause of death post-2011, while deaths attributed to mental disorders decreased significantly.
- Other categories, including self-poisoning, undetermined deaths, and assault, exhibit less consistent trends, highlighting the multifaceted nature of drug-related mortality in Scotland.

# Create the faceted ggplot with different scales for each cause of death
faceted_plot <- ggplot(table_2_long, aes(x = year, y = deaths, color = cause_of_death, group = cause_of_death)) +
  geom_line(size = 1) +  # Add lines for each cause, to visualize the trend over time
  geom_point(size = 2) +  # Add points for each year, to highlight individual data points
  geom_vline(xintercept = 2011, linetype = "dashed", color = "navy", size = 0.8) +  # Add policy change line at 2011, to indicate policy impact
  labs(
    title = "Drug Misuse Deaths Over Time by Cause (Pre- and Post-2011)",
    x = "Year",
    y = "Number of Deaths",
    color = "Cause of Death"
  ) +
  facet_wrap(~ cause_of_death, scales = "free_y") +  # Separate plots by cause with independent y-axes, to show trends for each cause separately
  theme_minimal() +
  theme(
    strip.text = element_text(size = 12, face = "bold"),  # Bold facet titles, make them stand out
    legend.position = "none",  # Remove legend, as it's redundant with facet labels
    plot.title = element_text(size = 16, face = "bold"),
    axis.title = element_text(size = 12),
    axis.text = element_text(size = 10)
  ) +
  aes(text = paste(
    "Year:", year,
    "<br>Deaths:", deaths,
    "<br>Cause of Death:", cause_of_death
  ))  # Add custom hover text, provides more details on hover

# Convert to interactive plotly object
interactive_faceted_plot <- ggplotly(faceted_plot, tooltip = "text") %>%
  layout(
    title = list(text = "Drug Misuse Deaths Over Time by Cause (Pre- and Post-2011)"),
    hoverlabel = list(font = list(size = 12))  # Adjust hover text size, to improve readability
  )

# View the interactive plot
interactive_faceted_plot

Self-poisoning and deaths of undetermined intent, also visualised in Figure 4, highlight the intersection between drug misuse and mental health crises. Scotland’s failure to integrate mental health and addiction services exacerbates these risks, leaving vulnerable individuals without adequate support.

Furthermore, there appears to be a growing role of poly-drug use, which complicates prevention efforts.
- Polydrug use contributes to over 70% of deaths (Figure 5), significantly increasing the risk of fatal interactions.
- This dominance of poly-drug use reveals shortcomings in current treatment models, which often address single-substance dependency rather than the interplay of multiple drugs (Ives and Ghelani, 2006).

---

Figure 5: Single-Drug vs. Poly-Drug Deaths in 2023

• Poly-drug deaths accounted for 73.7% (864) of total drug-related deaths, highlighting the complexity and heightened risks associated with substance interactions.
  
• Single-drug deaths represented 26.3% (308) of the total, underscoring the need for interventions targeting poly-drug use specifically.

# Load and clean the necessary data for single-drug and poly-drug deaths
table_8 <- read.xlsx(file_path, sheet = "Table_8", startRow = 5) %>%
  clean_names()

table_3 <- read.xlsx(file_path, sheet = "Table_3", startRow = 6) %>%
  clean_names()

# Filter data for 2023
single_drug_2023 <- table_8 %>%
  filter(!grepl("all drug misuse deaths", tolower(drugs_implicated), fixed = TRUE)) %>%  # Exclude total deaths, to focus on individual drugs
  summarize(single_drug_deaths = sum(persons, na.rm = TRUE)) %>%
  pull(single_drug_deaths)

total_deaths_2023 <- table_3 %>%
  filter(year == 2023) %>%
  pull(all_drug_misuse_deaths)

# Calculate poly-drug deaths by subtracting single-drug deaths from total deaths
poly_drug_2023 <- total_deaths_2023 - single_drug_2023

# Create a summary data frame for the plot
drug_summary_2023 <- data.frame(
  type = c("Single-Drug Deaths", "Poly-Drug Deaths"),
  deaths = c(single_drug_2023, poly_drug_2023)
) %>%
  mutate(percentage = round(deaths / sum(deaths) * 100, 1))  # Calculate percentage of each type, to show proportion of deaths

# Plot the data as a bar chart comparing single-drug and poly-drug deaths
poly_vs_single_plot <- ggplot(drug_summary_2023, aes(x = deaths, y = type, fill = type)) +
  geom_bar(stat = "identity") +  # Create bar chart, comparing deaths between the two types
  geom_text(aes(
    label = paste0(type, "\nDeaths: ", deaths, "\nPercentage: ", percentage, "%"),
    x = deaths / 2  # Position text in the center of each bar, to display death count and percentage
  ),
  color = "white",    # Use white text for contrast, to make it visible against the bars
  size = 5,           # Set font size, ensuring text is readable
  hjust = 0.5         # Center text horizontally, improves text alignment
  ) +
  labs(
    title = "Single-Drug vs Poly-Drug Deaths (2023)",
    x = "Number of Deaths",
    y = NULL,          # Remove Y-axis label, no need for it in this plot
    fill = NULL        # Remove fill legend, it's unnecessary here
  ) +
  theme_minimal() +
  theme(
    axis.text.y = element_blank(),        # Remove Y-axis text, to clean up the plot
    axis.ticks.y = element_blank(),       # Remove Y-axis ticks, to clean up the plot
    legend.position = "none",             # Remove legend, as mot needed for this chart
    plot.title = element_text(size = 16, face = "bold", hjust = 0.5),  # Style and center title, making it stand out
    axis.title.x = element_text(size = 12)  # Set X-axis label size, to improve readability
  ) +
  scale_fill_brewer(palette = "Set2")  # Apply a color palette, to make the plot more visually appealing

poly_vs_single_plot

Addressing these causes requires interventions that extend beyond harm reduction.

• The widespread use of methadone for opioid dependency, dispensed by 79.1% of pharmacies in 2006, demonstrates a commitment to harm reduction.
  
• However, the lack of progress in needle exchange services remains a significant gap in addressing drug misuse effectively.

---

Socioeconomic Disparities

Socioeconomic deprivation is a key driver of drug misuse deaths in Scotland:

• Individuals in the most deprived quintile (SIMD 1) experience death rates up to eight times higher than those in the least deprived quintile (SIMD 5).
  
• Factors such as the affordability and accessibility of ‘street benzos’ in deprived areas exacerbate this crisis (McAuley et al., 2022).

Addressing these disparities requires structural reforms to ensure:
1. Equitable funding for harm reduction services.
2. Improved access to essential interventions in the most deprived areas.

---

Figure 6: Drug Misuse Deaths Over Time by SIMD Quintile (2000–2023)

• Age-standardized rates highlight stark disparities:
  • Individuals in the most deprived quintile (SIMD 1) consistently experience the highest rates of drug misuse deaths.
    
  • In contrast, the least deprived quintile (SIMD 5) shows significantly lower rates, underscoring the impact of socioeconomic inequalities.

# Load and clean Table 9 (Quintiles): Read and clean the table containing quintile data
table_9 <- read.xlsx(file_path, sheet = "Table_9", startRow = 5) %>%
  clean_names()

# Reshape the data: Pivot data to long format and clean quintile names
quintile_data <- table_9 %>%
  select(year, starts_with("simd_quintile")) %>%  # Select relevant columns, to focus on quintiles
  pivot_longer(cols = -year, names_to = "quintile", values_to = "age_standardized_rate") %>%  # Pivot to long format, simplifies data for analysis
  mutate(quintile = gsub("simd_quintile_", "", quintile))  # Clean quintile names, to remove prefix for clarity

# Clean the data: Filter and label quintiles, select relevant columns
quintile_data_clean <- quintile_data %>%
  filter(grepl("age_standardised_rate", quintile)) %>%  # Filter for age-standardized rates, to focus on relevant data
  mutate(
    quintile_number = case_when(  # Label quintiles, to clarify most and least deprived
      grepl("1_most_deprived", quintile) ~ "1 (Most Deprived)",
      grepl("5_least_deprived", quintile) ~ "5 (Least Deprived)",
      TRUE ~ gsub("_age_standardised_rate", "", quintile)
    )
  ) %>%
  select(year, quintile_number, age_standardized_rate)  # Select relevant columns, keeping only necessary data

# Check for missing values in age_standardized_rate
quintile_data_clean %>% filter(is.na(age_standardized_rate))

# Create Quintile Line plot: Visualize trends over time for different quintiles
quintile_plot <- ggplot(quintile_data_clean, aes(x = year, y = age_standardized_rate, color = quintile_number, group = quintile_number)) +
  geom_line(size = 1.2) +  # Plot lines for quintiles, to show trends for each quintile
  geom_point(size = 1.5, alpha = 0.8) +  # Add points for data clarity, to highlight individual data points
  labs(
    title = "Drug Misuse Deaths Over Time by SIMD Quintile",
    subtitle = "Highlighting disparities in age-standardized rates (2000–2023).",
    x = "Year",
    y = "Age-Standardized Rate (per 100,000)",
    color = "Quintile"
  ) +
  theme_minimal() +
  theme(
    legend.position = "right",  # Position legend on the right, clears presentation
    legend.title = element_text(size = 10),
    legend.text = element_text(size = 9)
  ) +
  scale_color_brewer(palette = "Set2") +  # Apply color palette, to make the plot visually appealing
  aes(text = paste("Year:", year, "<br>Age-Standardized Rate:", round(age_standardized_rate, 2), "<br>Quintile:", quintile_number))  # Add custom hover text, which provides detailed info on hover

# Convert to interactive plot: Make the plot interactive with Plotly
quintile_interactive_plot <- ggplotly(quintile_plot, tooltip = "text")  # Add interactivity, allowing users to explore data by hovering over points

# View Quintile Interactive Plot
quintile_interactive_plot

Key Takeaway:
Addressing the stark socioeconomic disparities requires prioritizing harm reduction services in the most deprived communities. Policies should specifically target barriers to access, including:
- Affordability of services.
- Stigma associated with drug misuse.
- Equitable distribution of harm reduction services.

---

The funding formula used by NHS Scotland Boards and Alcohol and Drug Partnerships (ADPs) has disproportionately impacted deprived communities, leading to:
- Increased Drug-Related Death (DRD) rates (McPhee and Sheridan, 2020).
- Limited participation of pharmacies in needle exchange programs due to:
- Safety concerns.
- Perceived impact on customer relationships.

Despite modest improvements, only 12.5% of Scottish pharmacies provided needle exchange services as of 2006, reflecting the persistent challenges in achieving equitable service distribution (Matheson, Bond, and Tinelli, 2007).

---

Figure 7: Drug Misuse Deaths Over Time by SIMD Decile (2000–2023)

• Age-standardized rates reveal stark inequalities:
  • The most deprived decile (1) consistently experiences the highest rates of drug misuse deaths, peaking in 2020.
    
  • In contrast, the least deprived decile (10) shows consistently low rates throughout the period.
    
  • These trends highlight the socioeconomic gradient in drug misuse mortality and the urgent need for targeted interventions.

# Load and clean Table 10 (Deciles): Read and clean decile data from the table
table_10 <- read.xlsx(file_path, sheet = "Table_10", startRow = 6) %>%
  clean_names()

# Process the decile data in a single step: Select relevant columns, clean and reshape the data
decile_data <- table_10 %>%
  select(year, starts_with("simd_decile")) %>%  # Select columns for deciles, to focus on relevant data
  mutate(across(starts_with("simd_decile"), ~ as.numeric(gsub("[^0-9.]", "", .)))) %>%  # Clean non-numeric values, ensuring data is numeric for analysis
  pivot_longer(cols = -year, names_to = "decile", values_to = "age_standardized_rate") %>%  # Pivot to long format, to simplify data for analysis
  mutate(decile = gsub("simd_decile_", "", decile))  # Clean up column names, to make them more readable

# Clean and filter decile data: Separate decile values, clean up labels, and filter out missing values
decile_data_clean <- decile_data %>%
  separate(decile, into = c("decile_number", "metric"), sep = "_(?=[^_]+$)", extra = "merge", fill = "right") %>%  # Separate decile info, to split into more meaningful columns
  filter(metric == "rate") %>%  # Filter for age-standardized rates, to focus on relevant data
  mutate(
    decile_number = gsub("_most_deprived", "", decile_number),  # Clean decile labels, to remove unnecessary text
    decile_number = gsub("_age_standardised", "", decile_number),
    decile_number = gsub("_least_deprived", "", decile_number),
    decile_number = as.factor(decile_number),  # Convert decile to a factor, to make it easier to categorize
    decile_label = case_when(  # Label deciles, making them more readable
      decile_number == "1" ~ "1 (Most Deprived)",
      decile_number == "10" ~ "10 (Least Deprived)",
      TRUE ~ as.character(decile_number)
    )
  ) %>%
  filter(!is.na(age_standardized_rate))  # Filter out rows with missing data, to clean the data for analysis

# Assign colors for the deciles
decile_colors <- c(RColorBrewer::brewer.pal(8, "Set2"), "#Bd82ff", "#A0c4ff")  # Set color palette, to make the plot visually appealing

# Create the decile plot: Plot the age-standardized rate over time by decile
decile_plot <- ggplot(decile_data_clean, aes(x = year, y = age_standardized_rate, color = decile_number, group = decile_number)) +
  geom_line(data = decile_data_clean %>% filter(decile_number %in% c("1", "10")), aes(color = decile_number), size = 1.2) +  # Highlight most and least deprived, emphasizing key quintiles
  geom_line(data = decile_data_clean %>% filter(!decile_number %in% c("1", "10")), aes(color = decile_number), alpha = 0.7, size = 0.7) +  # Plot other deciles, to show all trends
  geom_point(size = 1.5, alpha = 0.6) +  # Add points for clarity, to highlight data points on the line
  labs(
    title = "Drug Misuse Deaths Over Time by SIMD Decile",
    subtitle = "Highlighting disparities in age-standardized rates (2000–2023).",
    x = "Year",
    y = "Age-Standardized Rate (per 100,000)",
    color = "SIMD Decile"
  ) +
  theme_minimal(base_size = 12) +
  theme(
    plot.title = element_text(size = 14, face = "bold"),  # Style plot title, to make it stand out
    plot.subtitle = element_text(size = 12, face = "italic"),  # Style subtitle,to make it visually appealing
    legend.title = element_text(size = 11),  # Style legend title, improves readability
    legend.text = element_text(size = 10)  # Style legend text, improves readability
  ) +
  scale_color_manual(values = decile_colors, name = "SIMD Decile", breaks = 1:10, labels = c("1 (Most Deprived)", "2", "3", "4", "5", "6", "7", "8", "9", "10 (Least Deprived)")) +  # Add color scale, improves visual clarity
  aes(text = paste("Year:", year, "<br>Age-Standardized Rate:", round(age_standardized_rate, 2), "<br>Decile:", decile_label))  # Add custom hover text, provides detailed info on hover

# Convert to interactive Plotly plot
decile_interactive_plot <- ggplotly(decile_plot, tooltip = "text")  # Make the plot interactive, allows user interactivity

# View the interactive plot
decile_interactive_plot

Policy responses have been inadequate in addressing the structural drivers of drug misuse. Critiques of the governance of drug and alcohol services since 2009 emphasize the following issues:
- Widened inequalities due to funding decisions.
- Disproportionate risks for already vulnerable communities (McPhee and Sheridan, 2020).

Governance failures undermine efforts to tackle the systemic drivers of drug misuse and drug-related deaths (DRDs). Key challenges include:
- Concentration of needle exchange services in urban areas, limited further by:
- Stigma.
- Resource constraints (Matheson, Bond, and Tinelli, 2007) (Childs et al., 2021).

Addressing these disparities requires:
- Investments in affordable housing.
- Initiatives for job creation.
- Development of integrated, community-based health services.

---

Geographic Patterns

Geographic differences in drug misuse reflect Scotland’s uneven distribution of social and health resources:

• Rural areas, which accounted for only 18.1% of surveyed pharmacies in 2006, face significant limitations in harm reduction service availability.
  • This underscores the need for targeted interventions to bridge geographic disparities in service provision (Matheson, Bond, and Tinelli, 2007).
• Urban areas like Glasgow experience higher rates of drug deaths due to the intersection of:
  • Homelessness.
    
  • Poverty.
    
  • Social exclusion (Figure 8).
• In contrast, rural areas report lower rates but encounter challenges like:
  • Geographic isolation.
    
  • Limited harm reduction services (Bailey, Bramley, and Gannon, 2016).

---

Figure 8: Total Drug Misuse Deaths by Health Board Area in Scotland (2010–2023)

• The map illustrates geographic disparities in drug misuse deaths:
  • Urban areas, particularly Greater Glasgow and Clyde, exhibit the highest concentration of deaths, shown in dark blue.
    
  • Rural and less densely populated regions report significantly lower totals.
    
  • These trends emphasize regional inequalities in drug misuse mortality.

# Import health board shapefile: Read the shapefile containing health board boundaries
hb_shapes <- st_read("NHS_healthboards_2019.shp") %>% 
  clean_names()

# Join health board shape data with death data: Read death data and filter for "Persons"
table_HB1 <- read.xlsx(file_path, sheet = "Table_HB1", startRow = 5)

persons_hb1 <- table_HB1 %>% 
  filter(Sex == "Persons") %>%  # Filter for person-level data, to focus on total deaths
  select(-Scotland)  # Remove 'Scotland' column, its not relevant for region-specific analysis

# Reshape the data: Pivot health board data into long format
long_HB1 <- persons_hb1 %>% 
  pivot_longer(
    cols = c(Ayrshire.and.Arran:Western.Isles), 
    names_to = "hb_name", 
    values_to = "deaths" 
  )

# Recode health board names for clarity
long_HB1 <- long_HB1 %>%
  mutate(hb_name = recode(hb_name,
    "Ayrshire.and.Arran" = "Ayrshire and Arran", 
    "Dumfries.and.Galloway" = "Dumfries and Galloway", 
    "Greater.Glasgow.and.Clyde" = "Greater Glasgow and Clyde", 
    "Forth.Valley" = "Forth Valley", 
    "Western.Isles" = "Western Isles"
  ))

# Join health board deaths data with shape data
hb_deaths <- long_HB1 %>% 
  full_join(hb_shapes, by = "hb_name")

# Ensure data is in the correct format for plotting (sf object)
hb_deaths_sf <- st_as_sf(hb_deaths)

# Summarize total deaths by health board
total_deaths_by_region <- hb_deaths_sf %>%
  group_by(hb_name) %>%
  summarize(total_deaths = sum(deaths, na.rm = TRUE))

# Transform the data to WGS84 for leaflet compatibility
hb_deaths_sf <- st_transform(hb_deaths_sf, crs = 4326)

# Summarize total deaths by region (Health Board)
total_deaths_by_region <- hb_deaths_sf %>%
  group_by(hb_name) %>%
  summarize(total_deaths = sum(deaths, na.rm = TRUE), .groups = "drop")

# Create the interactive leaflet map: Plot health boards with drug misuse death data
hb_map <- leaflet(total_deaths_by_region) %>%
  addTiles() %>%  # Add base map tiles, to show geographic context
  addPolygons(  # Add polygons for each health board, to represent regions on the map
    fillColor = ~colorNumeric("YlOrRd", total_deaths)(total_deaths),  # Set fill color based on total deaths, to visualize death rates
    weight = 1,
    opacity = 1,
    color = "white",
    fillOpacity = 0.7,
    popup = ~paste(
      "<strong>Health Board:</strong>", hb_name,
      "<br><strong>Total Deaths:</strong>", total_deaths
    )
  ) %>%
  addLegend(  # Add legend for color scale, to help users interpret map colors
    pal = colorNumeric("YlOrRd", total_deaths_by_region$total_deaths),
    values = total_deaths_by_region$total_deaths,
    title = "<b>Total Deaths</b>",
    position = "bottomright"
  ) %>%
  addScaleBar(  # Add scale bar, to provide distance context on the map
    position = "bottomleft",
    options = scaleBarOptions(imperial = FALSE)
  ) %>%
  addControl(  # Add custom control with map title, to provide description for users
    HTML("<div style='background-color: white; padding: 10px; border-radius: 5px; box-shadow: 0px 0px 5px gray;'>
            <h3 style='margin: 0;'>Interactive Map of Drug Misuse Deaths in Scotland (2010–2023)</h3>
          </div>"),
    position = "topright"
  ) %>%
  addControl(  # Add note for context, explains the data shown
    HTML("<div style='background-color: white; padding: 10px; border-radius: 5px; box-shadow: 0px 0px 5px gray;'>
            <p style='margin: 0;'>Note: Data reflects total drug misuse deaths reported by region from 2010 to 2023.</p>
          </div>"),
    position = "bottomleft"
  )

hb_map

Expanding harm reduction services in high-risk regions like Glasgow, such as mobile units or supervised injection sites, can help reduce the geographic disparity in drug-related deaths.

Critically, the allocation of resources has not aligned with regional needs. Due to Glasgow’s high mortality rates (Figure 8), funding for harm reduction services in the area has been insufficient to meet demand (Tweed et al., 2018).
- The expansion of mobile harm reduction units and supervised injection facilities could address some of these disparities.
- By doing so, Scotland would be one step closer to ensuring its vulnerable populations have access to life-saving interventions.

---

Poly-Drug Use vs. Single-Drug Use

• Over 70% of drug-related deaths in 2023 were linked to poly-drug use (Figure 5).
  
• Current treatment models, which focus on single-substance dependencies, are inadequate.

Key Issues and Risks

1. Poly-drug users frequently use benzodiazepines, particularly etizolam, which is implicated in fatal overdoses.
   • Only 1% of deaths involved etizolam alone (McAuley et al., 2022).
     
2. The combination of substances, especially benzodiazepines and opioids, dramatically increases the risk of fatal overdose (Boon et al., 2020).

Interventions Needed

• Expand detoxification and rehabilitation programs that address interactions between multiple substances.
  
• Improve education on the risks of poly-drug use.
  • Distribute naloxone alongside targeted public health campaigns to warn of the risks associated with combining substances.

---

Policy Implications and Recommendations

Expanding Harm Reduction Services

• Invest in supervised injection sites, mobile harm reduction units, and needle exchange programs.

Addressing Systemic Inequities

• Implement equitable funding for deprived regions.
  
• Integrate mental health and addiction services.

Targeted Public Health Campaigns

• Promote education on the risks of poly-drug use.
  
• Expand naloxone distribution programs and training initiatives.

---

Conclusion

Scotland’s drug misuse deaths result from a complex interplay of socioeconomic, behavioural, and policy failures. Over the past three decades, systemic shortcomings have fuelled one of Europe’s highest drug-related death rates.

Key Challenges

1. Persistent poverty and untreated mental health crises.
   
2. Inadequate harm reduction services.
   
3. Governance failures and inequitable funding practices (McPhee and Sheridan, 2020).

Recommendations

• Sustained investment in harm reduction strategies.
  
• Structural reforms to ensure equitable access to services.
  
• Integration of mental health and addiction care.

The success of the National Naloxone Programme demonstrates the potential of targeted, well-funded interventions to reduce high-risk mortality (Bird et al., 2016).
- However, such efforts must be scaled and complemented by broader public health initiatives that prioritise prevention, equity, and long-term community resilience.

By adopting a proactive, equity-focused approach, Scotland can pave the way for lasting systemic change, ultimately saving lives and improving community resilience.

---

Sources and References

Data Source

• National Records of Scotland (2023) Drug-related deaths in Scotland in 2023. Available at: National Records of Scotland

References

1. Bailey, N., Bramley, G. and Gannon, M., 2016. Poverty and social exclusion in urban and rural areas of Scotland
2. Bird, S.M., McAuley, A., Perry, S. and Hunter, C., 2016. Effectiveness of Scotland’s National Naloxone Programme for reducing opioid‐related deaths: a before (2006–10) versus after (2011–13) comparison. Addiction, 111(5), pp.883-891.
3. Boon, M., van Dorp, E., Broens, S. and Overdyk, F., 2020. Combining opioids and benzodiazepines: effects on mortality and severe adverse respiratory events. Annals of Palliative Medicine, 9(2), pp.54257-54557.
4. Childs, E., Biello, K.B., Valente, P.K., Salhaney, P., Biancarelli, D.L., Olson, J., Earlywine, J.J., Marshall, B.D. and Bazzi, A.R., 2021. Implementing harm reduction in non-urban communities affected by opioids and polysubstance use: a qualitative study exploring challenges and mitigating strategies. International Journal of Drug Policy, 90, p.103080.
5. Gjersing, L. and Helle, M.K., 2021. Injecting alone is more common among men, frequent injectors and polysubstance users in a sample of people who inject drugs. Substance Use & Misuse, 56(14), pp.2214-2220.
6. Gossop, M., Griffiths, P. and Strang, J., 1994. Sex differences in patterns of drug taking behaviour: A study at a London community drug team. The British Journal of Psychiatry, 164(1), pp.101-104.
7. Ives, R. and Ghelani, P., 2006. Polydrug use (the use of drugs in combination): A brief review. Drugs: education, prevention and policy, 13(3), pp.225-232.
8. Matheson, C., Bond, C.M. and Tinelli, M., 2007. Community pharmacy harm reduction services for drug misusers: national service delivery and professional attitude development over a decade in Scotland. Journal of public health, 29(4), pp.350-357.
9. Matheson, C., Pflanz-Sinclair, C., Aucott, L., Wilson, P., Watson, R., Malloy, S., Dickie, E. and McAuley, A., 2014. Reducing drug related deaths: a pre-implementation assessment of knowledge, barriers and enablers for naloxone distribution through general practice. BMC family practice, 15, pp.1-10.
10. McAuley, A., Matheson, C. and Robertson, J.R., 2022. From the clinic to the street: the changing role of benzodiazepines in the Scottish overdose epidemic. International Journal of Drug Policy, 100, p.103512.
11. McPhee, I. and Sheridan, B., 2020. AUDIT Scotland 10 years on: Explaining how funding decisions link to increased risk for drug related deaths among the poor. Drugs and Alcohol Today, 20(4), pp.313-322.
12. O’sullivan, R., Burns, A., Leavey, G., Leroi, I., Burholt, V., Lubben, J., Holt-Lunstad, J., Victor, C., Lawlor, B., Vilar-Compte, M. and Perissinotto, C.M., 2021. Impact of the COVID-19 pandemic on loneliness and social isolation: A multi-country study. International journal of environmental research and public health, 18(19), p.9982.
13. Thompson, R.A., Sanderson, W.T., Westneat, S., Bunn, T., Lavender, A., Tran, A., Holsinger, C., Flammia, D., Zhang, L. and He, Y., 2021. Perceptions of opioid and other illicit drug exposure reported among first responders in the southeast, 2017 to 2018. Health Science Reports, 4(3), p.e335.
14. Tweed, E.J., Miller, R.G., Schofield, J., Barnsdale, L. and Matheson, C., 2022. Why are drug-related deaths among women increasing in Scotland? A mixed-methods analysis of possible explanations. Drugs: Education, Prevention and Policy, 29(1), pp.62-75.
15. Tweed, E.J., Rodgers, M., Priyadarshi, S. and Crighton, E., 2018. “Taking away the chaos”: a health needs assessment for people who inject drugs in public places in Glasgow, Scotland. BMC public health, 18, pp.1-9.